Advanced Computer Vision Techniques for Marine Ecology

海洋生态学的先进计算机视觉技术

• 批准号: DP150104251
• 负责人: Prof Dr Mohammed Bennamoun
• 金额: $ 43.04万
• 依托单位: The University of Western Australia
• 依托单位国家: 澳大利亚
• 项目类别: Discovery Projects
• 财政年份: 2015
• 资助国家: 澳大利亚
• 起止时间: 2015-01-01 至 2021-12-30
• 项目状态: 已结题
• 来源: https://dataportal.arc.gov.au/RGS/Web/Grants/DP150104251
• 关键词: Advanced; Computer; Vision; Techniques; Marine

Ever expanding human activity coupled with climate change has severely damaged marine ecosystems, which play a key role in our planet's ability to sustain life. Yet automated technology to monitor the health of our oceans still does not exist, with marine scientists still having today to process manually a massive amount of raw underwater imagery. This research aims to address this bottleneck by developing advanced computer vision tools for rapid, large-scale, automatic identification of marine species. Such an automated technology is expected to greatly benefit marine ecological studies in terms of speed, cost, accuracy of the spatial/temporal sampling and thus in better quantifying the level of environmental change marine ecosystems can tolerate.

不断扩大的人类活动加上气候变化严重破坏了海洋生态系统，而海洋生态系统在地球维持生命的能力中发挥着关键作用。然而，监测海洋健康状况的自动化技术仍然不存在，海洋科学家今天仍然需要手动处理大量的原始水下图像。本研究旨在通过开发先进的计算机视觉工具来快速、大规模、自动识别海洋物种，从而解决这一瓶颈。预计这种自动化技术将在速度、成本、时空采样的准确性方面大大有利于海洋生态研究，从而更好地量化海洋生态系统所能容忍的环境变化水平。